The gendered effects of entrepreneurialism in contrasting contexts

Purpose – Contrasting Sweden and Tanzania, this paper aims to explore the experiences of women entrepreneursaffected by entrepreneurialism. This study discusses the impact on their position in society and ontheir ability to take feministaction. Design/methodology/approach – This paper analysed interviews conducted in the two countries over 15years,usinga holistic perspective oncontext, including itsgendered dimensions. Findings – The results amount to a critique of entrepreneurialism. Women in Sweden did not experience much gain from entrepreneurship, while in Tanzania results were mixed. Entrepreneurialism seems unable to improve the situation for women in the relatively well-functioning economies in the global north, where it was designed. Research limitations/implications – In mainstream entrepreneurship studies, there is a focus on the institutional context. From the analysis, it is apparent that equal attention must be given to the social and spatial contexts, as they may have severe material and economic consequences for entrepreneurs and entrepreneurship. The paper raises questions for further studies on the gendering of markets in different contexts, aswell as questions ontheurban-rural dimension. Practical implications – In Sweden, marketisation of welfare services led to more women-owned businesses, but the position of women did not improve. The results strongly convey the need for a careful analysis ofthe pre-existingcontext,beforeinitiatingreforms. Originality/value


Introduction
Entrepreneurialism is part and parcel of neo-liberal policy, which implies new forms of political-economic governance premised on the extension of market relationships (Larner, 2000).Entrepreneurialism sees entrepreneurshipin the narrow sense of start-ups and SMEsas the solution to a large variety of problems (Harvey, 1989), but how it affects the position of women in society has been a matter of debate.Proponents talk about results such as more viable economies and increased independence for women (Brush et al., 2006;Allen et al., 2008).Critics say that the reliance on entrepreneurship in lieu of public sector jobs and a tax-funded safety net has the opposite effect (Sundin, 2011;Kantola and Squires, 2012;Ahl and Marlow, 2019;Ahl and Nelson, 2015).The cited critics have, however, studied effects of neo-liberal reforms in industrialised, developed states.Like others before us (Welter, 2011;Gaddefors and Anderson, 2017) we argue that analysis of such effects must take the context into account.To add to our knowledge of how outcomes of entrepreneurialism depend on context, this paper aims to explore the experiences of women entrepreneurs in two different and highly contrasting countries: Sweden, known for gender equality and an encompassing welfare state and Tanzania, where gender equality is low, and the state is weak.Indeed, entrepreneurship studies, in general, need to be extended from the European and North American context (Jaim and Islam, 2018;Mazonde and Carmichael, 2016;Palali c et al., 2020).
Both Tanzania and Sweden have been deeply affected by neo-liberal agendas, including entrepreneurialism, but in different ways.In Tanzania, we have seen the state and also the donor countries conform to the neo-liberal "Washington Consensus", which includes the liberalisation of domestic markets through privatisation and deregulation.In Sweden, we have witnessed the adoption of New Public Management (Hood, 1991) and a movement away from the "Scandinavian welfare state regime" (Esping-Andersen, 1990).A previous study (Authors, forthcoming) outlined and contrasted the different economic reforms that resulted from the implementation of the neo-liberal agenda in Sweden and Tanzania.The study proposed that neo-liberal agendas would have negative effects on the position of women in society in Sweden and more mixed effects in Tanzania.The main reason for the proposed difference was the pre-existing institutions, for example, the women-friendly welfare state in Sweden and the absence of such in Tanzania.
In this paper, we explore the experiences of women entrepreneurs affected by entrepreneurialism and discuss the impact on their position in society and on their ability to take feminist action, i.e. to improve the situation for themselves and for other women.We do so by revisiting over 70 interviews with women entrepreneurs conducted during the 15-year time period of neo-liberal reforms in the two countries.Methodologically, the study is interpretative and inductively driven (Dana and Dana, 2005), and we use contrasting as an analytical approach (Marcus, 1986;Dana, 1997).In analysing the interviews, we use a holistic perspective on context, including its gendered dimensions (Welter et al., 2014).
The article concludes, in line with the propositions resulting from the literature study (Authors, forthcoming), that women in Sweden did not experience much gain from entrepreneurialism, while results were more mixed in the case of Tanzania.Women in Tanzania reported gaining economic independence from men and opportunities to support other women, but resilient, patriarchaleven abusivegender regimes were still recreated.From a feminist perspective, the results amount to a critique of entrepreneurialism.
Ironically, entrepreneurialism seems unable to improve the situation for women in the relatively well-functioning economies in the global north, where neo-liberalism was designed.This was particularly the case in a Scandinavian welfare state such as Sweden.Only in a context characterised by a weak state, high levels of corruption and unbridled patriarchy did neo-liberalism benefit (some) women entrepreneurs.With regard to policy, our results strongly convey the need for a careful analysis of the pre-existing context, before initiating reforms.
The article is structured as follows.Below, we outline our frame of reference in terms of the contextualisation of women entrepreneurs.Then follows a description of our methodology, before the empirical section, which is followed by discussion.The paper ends with our conclusions with regard to theoretical developments, policy implications and avenues for further research.

Contextualising women entrepreneurs
Entrepreneurship scholars have increasingly called for the need to attend to context (Steyaert and Hjorth, 2006;Welter, 2011;Dana and Ramadani, 2015), not least its gender dimensions (Ahl, 2006;Jennings and McDougald, 2007).Calls have also been made for extending the scope of research to areas other than developed nations in North America and Europe (Salamzadeh and Kawamorita Kesim, 2017;Jaim and Islam, 2018).Building on the framework and previous studies outlined below, this article contributes to that endeavour.

Definitions and framework
Entrepreneurship, broadly defined as enabling social change (Steyaert and Hjorth, 2006;Cal as et al., 2009), is a phenomenon that is affected by but also affects, the context.This complexity has been termed "the paradox of embedded agency" (Battilana, 2006) or the "double sociability of entrepreneurship" (Steyaert and Hjorth, 2006).Scholars have long recognised the contextual embeddedness of entrepreneurship (Granovetter, 1985).In accordance with Johns (2006), we understand context as situational opportunities and constraints that affect behaviour.The dimensions of context are multiple and multi-faceted, which is why Welter (2011) argues that an interdisciplinary perspective is needed.Drawing on Whetten (2009Whetten ( , 1989)), she distinguishes between the where context (business, social, spatial and institutional) and the when context (historical and temporal).
Within the field of women's entrepreneurship studies, many steps have been taken in the directions suggested by Ahl (2006).Research avoids essentialist assumptions and recognises the impact of context.Special issues and books in the field have focussed on context (Díaz-García et al., 2016;Yousafzi et al., 2018;Knezovi c et al., 2020;Achtenhagen and Tillmar, 2013).Risks of US-centric approaches and comparisons without consideration of contextual differences have been brought to attention (Marlow et al., 2008;Dana and Ramadani, 2015).
Context studies on women's entrepreneurship include studies on the social context in terms of household and family context (Jennings and McDougald, 2007) and everydayness (Rehn and Taalas, 2004;Welter and Smallbone, 2011).Drawing on the framework outlined by Welter (2011), Welter et al. (2014) argue that the gender dimension is central not only within the social context but also within the institutional, spatial and temporal contexts.By "gender" we refer to social constructions.In this article, gendered dimensions of context imply dimensions of context that affect men and women differently.We take a feminist perspective, which is broadly defined as the recognition of women's subordination and the desire to rectify this.

Gendered effects
Referring to geographical and feminist studies, Welter et al. (2014) re-conceptualise the spatial context into encompassing physical space (location) and place (norms and values dominating the particular space).The authors also highlight women's agency in changing their spatial and institutional context while at the same time problematising the argument that entrepreneurship empowers women.To quote Al-Dajani and Marlow (2013), "we are in danger of celebrating entrepreneurial activities as an idealised solution to poverty, marginalisation and subordination".Careful contextual analysis is called for, to understand the pre-conditions and possible "gender entrepreneurship gaps" (Piacentini, 2013) that may hinder the empowerment of women through entrepreneurship.Both interview studies in Zimbabwe (Mazonde and Carmichael, 2016) and Tanzania (Langevang et al., 2018) and a quantitative survey of Nigerian women entrepreneurs (Sajuyigbe and Fadeyibi, 2017) found socio-cultural context, work-home conflict and other gender-oriented factors hindering such empowerment.Welter et al. (2014) recognise the need to understand under what circumstances women experience the different dimensions of context as enabling or constraining for their agency.In this paper, we extend the framework and ask under what circumstances contextual dimensions also enable feminist agency, i.e. action that improves women's situation in society on a level beyond the single individual.Feminist agency through enterprise was coined "FemInc.ism"by Ahl et al. (2016), based on cases in which women had used the business form to enable institutional change (Tillmar, 2009;Berglund and Johansson, 2007).The concept FemInc.ismcan be used to analyse the potential for men and women to use the organisational form of enterprise and markets not only to enable change but also to analyse potential pitfalls and constraints on this route.

Women's entrepreneurship in the studied contexts
Since Scandinavian research on women's entrepreneurship has often focussed on the institutional context, including the gender order (Ahl and Nelson, 2015;Sundin and Tillmar, 2010;Sköld and Tillmar, 2015;Pettersson et al., 2017;Sundin, 2011), there is a body of knowledge on how women's entrepreneurship has been affected by neo-liberal reforms, particularly by the privatisation of former public services.As many entrepreneurs start businesses in their former profession (Shane, 2000) and since public welfare is heavily female-dominated in these countries, the number of women entrepreneurs was expected to increase when public welfare was privatised (Blomberg et al., 2011).However, things did not turn out the way neo-liberal politicians had expected; the outcomes were heavily gendered and dependent on class (Sundin, 2011).Most of the business opportunities on publicly funded markets for welfare services went to male-owned businesses, according to database studies on national level (Sköld, 2015;Sköld and Tillmar, 2015), as well as case studies on the local level (Sundin and Rapp, 2006;Sundin and Tillmar, 2010).The result has been explained by the resilience of the gender order, which was reproduced in the business context.The impact of the gender order on self-employed women, their incomes and career choices also in western countries is neither unknown (Hundley, 2000;Marshall and Flaig, 2014;Thébaud, 2016;Klyver et al., 2013), nor given adequate attention in mainstream entrepreneurship studies and policy-making (Pettersson et al., 2017).
In the East African, and hence Tanzanian, context, the view of women's entrepreneurship as the "untapped resource" for economic development (Ramadani et al., 2013;Ratten and Jones, 2018) still holds strong among development agencies.There is a long list of initiatives promoting women's entrepreneurship and market-driven development (Vossenberg, 2013) [1].Yet self-employed women as a group belong to the poorest of the poor (Tindiwensi, 2007).Among the recognised constraints are a heavy domestic workload (Kibera and Kibera, 1999), lack of social capital (cf.Komugisha Tindiwensi, 2007), lack of access to finance (Kibera and Kibera, 1999) and lack of access to commercial justice (Ellis, 2007;Tillmar, 2016a).Studies suggest that many donors are over-optimistic about the anticipated gender equality resulting from neo-liberal initiatives.Unless indigenous values such as communal values are incorporated into entrepreneurship thinking (Ratten and Dana, 2017), the individualisation inherent in the business form risks having adverse effects on solidarity among women (Sigalla and Carney, 2012).Studies have shown that when women in a patriarchal context venture into entrepreneurship, there is a "work/home conflict" and a "lack of moral support from the family" (Sajuyigbe and Fadeyibi, 2017;Kibera and Kibera, 1999).Qualitative and ethnographically inspired studies have revealed such terms to be euphemisms for increased domestic violence because husbands feel threatened (Tillmar, 2016a;Vyas et al., 2015).The effects of entrepreneurialism for women in the East African context are still under debate and research results are inconclusive.Kinyanjui (2012) argues that women in the neo-liberal era redefined their investment in local-level development into self-organising social alliances that provided a livelihood for the women and confronted the patriarchal system in solidarity.There are also shining examples of women using gendered constraints to create business opportunities (Tillmar, 2016b) and other strategies to handle gendered constraints (Langevang et al., 2018).

Methodological approach
To understand the contextual dimensions, we follow Dana and Dana's (2005) call for entrepreneurship studies to use a qualitative and interpretative approach starting from empirical observations.The research question emerged by contrasting empirical results from the Tanzanian and Swedish context.We grew increasingly puzzled by what appeared to be divergent results in the two contexts and decided to systematically re-analyse interviews from previous studies, using contrasting as an analytical strategy (Dana, 1990(Dana, , 1997;;Tillmar, 2006).A cumulative understanding was built by contrasting the countries in an iterative process of re-analysing our data set, theories and previous studies.We aimed for a theory-method fit in our literature review (Gehman et al., 2018) and focussed on seminal work on context and entrepreneurship, including gender and previous studies on the two countries.The overall approach is hence not purely inductive, but best characterised as "abductive" (Alvesson and Sköldberg, 2000;Dana and Dana, 2005).
We make no claims regarding the comparability of the two countries in the more positivist sense or on a "surface-level" (Alvesson and Sköldberg, 2000)."Comprehensive research" like this (Dana and Dumez, 2015) serves to highlight mechanisms, challenge existing knowledge and redefine or improve, existing models (Whetten, 1989).More specifically, we improve an existing model on contextual dimensions affecting women entrepreneurs and challenge existing knowledge on entrepreneurialism.Contrasting different cases is powerful not only for providing important theoretical conclusions but also for uncovering social phenomena and for understanding the reasons for different outcomes (Ragin, 2014).A context is better understood when an alternative pattern is used to provide a contrast (Brislin, 1980;Stewart et al., 1994;Marcus and Fischer, 2014;Dana, 1997).Being able to contrast a country with a high level of gender equality with a country with a low level of gender equality has been an advantage in this study.

Interview material
We revisited primary data from our own studies consisting of unstructured and semistructured interviews collected by the first author during ethnographically inspired

Gendered effects
qualitative studies over a period of 15 years, between 1997 and 2012 (Appendix).This was a period when many neo-liberal reforms geared towards entrepreneurialism took place in the two countries.All the projects concerned gender and entrepreneurship/ SMEs.Hence, we had a pre-existing data set, which could help us answer previously unexplored research questions.
The 73 interviews in our data set had the character of free-flowing open-ended conversations, and the interviewer also noted non-verbal communication (Groenland and Dana, 2019).All the interviews addressed conditions for women entrepreneurs in the context at hand.Parts of the interviews were specific to the project within which they were conducted, as specified in Appendix.Interviewees were selected through purposive snow-ball sampling with initial assistance from local collaborating universities and business support organisations.We analysed 36 tape-recorded and transcribed interviews with women entrepreneurs affected by the implementation of neo-liberal policies in Sweden.The interviews were conducted in Swedish.The author resides in Sweden and is familiar with the context.In Tanzania, we used 37 interviews with women entrepreneurs.The first author lived in Tanzania for a year and a half and learned to speak the local language, which was vital for access to rich data.As interviewees in Tanzania were generally not comfortable with tape recorders, and interviews were conducted in noisy environments, detailed notes were taken.

Analytical procedure
We selected 10 interviews from each country for further structured in-depth analysis on the basis that they provided especially rich data with regard to the current research questions.The transcripts and notes from these 20 selected interviews comprised the subset of the original primary data which were subject to a qualitative secondary analysis (QSA) (Irwin, 2013;Gladstone et al., 2007).Appendix 2 gives an anonymised overview of these interviewees in terms of age and line and size of business.In both countries, the studies were conducted in urban areas and in female dominated sectors.Most Tanzanian firms were in handicraft, trade or food processing, and in Sweden they were in health and care sectors, which were directly affected by the privatisations.
As our data was collected with the same epistemological approach as used in this paper and, of course, within the contexts we discuss here, we avoid the major drawbacks of using secondary data (Irwin, 2013).Quality is also ensured by the "fit" between the new question and the original data, as the question arose from the primary data, as well as the participation of the researcher who conducted the original studies (Gladstone et al., 2007).Using commonly used methods of analysing qualitative data is yet another route for us to ensure quality.
The selected subset of data was subject to a thematic document study (Groenland and Dana, 2019) using the rigorous, structured approach to analysing qualitative material called the Gioia methodology (Gioia et al., 2013;Gehman et al., 2018).It enables a structured approach which also helps to identify and visualise the link between first-order concepts in the informants' terms (in our case quotes) and the aggregate theoretical dimensions, using second-order themes (Gioia et al., 2013).Presenting the first-order concepts as quotes gives the reader the illustration dosage (Glaser, 2003) needed to gain a sense of the empirical material underpinning the inductively driven theory development.The data structures which were generated by the analyses are presented in Table 1 (Tanzania) and Table 2 (Sweden).
Selected quotes were also used in the empirical descriptions below, for illustration purposes and for providing an understanding of how we arrived at the second-order themes 1st order concepts/ representative quotes 2nd order themes Aggregate dimension "It is difficult to sell at the market.The men come and get priority.Sometimes men can cut 10% off the price because they have big businesses" (Patricia) "Travelling is a problem if you are alone.When you come back, people have stolen and done nothing.
It's a problem for both women and men but there are differences in how it can be tackled.Some partners or workers are afraid of women, they will see that there is nothing you can do" (Anum) "You have to make a lot of effort to get respect!They do not think you are ok from the start" (Veronica) "And when I found the yoga, I was home.And then I wanted to pass on to others what I had been fortunate to receive -and to do that I had to start my own business" (Lisa) "There was far too little room for the woman and her individual needs" (Erika) Strong motivation to take social responsibility

Social change through entrepreneurship
Table 2.
Data structure Sweden JEC generated by the analysis (Siggelkow, 2007;Gioia et al., 2013).The overall patterns and conclusions reflect the interview material in total, that is, all 73 interviews.

Results of interview analysis
Behind each of the 20 interviews in our QSA, there is a life history of a woman who ventured into entrepreneurship and self-employment, with its joys, opportunities, sorrows and obstacles.When contrasting the 10 stories from Tanzania and the 10 stories from Sweden, we were surprised to find the sorrows and obstacles more common in the stories from Sweden.Narrating each of the stories is outside the scope of this paper, and we refer to the original studies listed in Appendix 1. Below, we present the results of our structured Gioia-inspired (Gioia et al., 2013) re-analysis of this selected subset of our data.

Women entrepreneurs in Tanzania
Gendered constraints in entrepreneurship and feminist agency through entrepreneurship emerge as aggregate dimensions in the analysis, as illustrated in the data structure in Table 1.The second-order themes generated represent different forms of gendered constraints and feminist agency, respectively.In Table 1, we have inserted representative quotes from the 10 selected interviews.In the text below, one quote is chosen to illustrate each second-order theme.

Gendered constraints in entrepreneurship.
There is no doubt that there are many gendered constraints to women's entrepreneurship in Tanzania.The entrepreneurs report that there is a strong gender bias among business contacts, affecting not only the price they can obtain on the market and the respect they are given within their own companies but also their ability to defend their rights in cases of theft or fraud.Some partners or workers are not afraid of women, they see that there is nothing you can do.Aumn.
The difficulty single women face in defending their business interests and rights against collaborators and employees is even higher.Fatima, a 45-year-old spice trader whose husband left her to live with his second wife when she pursued her business, had her property vandalised by neighbouring youngsters.Others narrate that without the support of a husband, business contacts can cheat them in many ways.
If you are not married, they are not afraid of you at all, they can do anything.Barika Gendered constraints could also come from within the household, in cases where the husbands were not supportive of the business.Intimate partner violence against women was another theme raised in the interviews.
I was in the batik business.He started to scream and fight.So I stopped that business and started with vegetables.The business grew a bit, then he started big fights and destroyed everything in the field and destroyed the vegetables.Mary 4.1.2Feminist agency through entrepreneurship.Many Tanzanian women perceived that membership in a Chamawomen's groupis vital for their ability to run their business.The enabling aspect of Chamas for women's entrepreneurship has previously been highlighted by Kinyanjui (2012).Within a Chama, women help each other practically, economically and socially/emotionally.A common example is a help they give each other to attend to the home and children when it is necessary to travel on business, as illustrated by the quote from Linda.In a Chama, it is common to have a savings and credit scheme where five to ten women take turns in giving and receiving funds; for example, everyone Gendered effects contributes a sum of money each week and they take turns to receive those funds (Table 1).At times, it is a way of saving and being able to make investments without interference from husbands.
We can help each other.To give each other information, for example if there is exhibitions somewhere everyone can get the information.The 'chama' is very lively.Barika I have started a group with people from here.We give each other advice and we have started to lend each other small sums of money.We meet every Saturday from 9 to 10.If you are a woman, you always have certain kinds of problems.Mary Hence, membership of a business-oriented Chama is a strong enabling factor for women entrepreneurs and for their agency, and the membership can itself be seen as a form of feminist agency.
What is enabled by entrepreneurship is also independence from individual men.Women's entrepreneurship not only sustains the livelihoods of the women themselves and their children but also funds a good education not only for sons but also for daughters.Patricia, a case in point, has a small shop in the outskirts of Dar es Salaam, practices smallscale food-processing, breeds chickens and keeps a few cattle from which she sells milk.She lives behind the shop with her children.It is afternoon and her daughter takes care of the shop-keeping during the interview.Patricia is very proud that she has been able to bring her children up by herself.

I have struggled to put my children [my daughter and her little brother] through [private] school right up to university. Patricia
A few of the interviews have explicit business goals to develop younger women and girls, and not only Chama members and daughters.Karen challenged the gender order when faced with the choice between her business and her husband (Author, 2016a).She chose the business, took out a divorce and now lives on the proceeds of a business in a male-labelled industry.Karen has had countless problems related to the institutional and social contextual dimensions, being a single woman in need of access to title deeds and proper contracts.Yet she is determined to carry on and to contract rural female farmers as her suppliers to empower them as well.Karen uses her business and coaching skills to support other women, especially those who are subject to intimate partner violence.I'm coaching women who keep poultry, backyard poultry.We see these problems.Karen Jasmine, an economically independent upper-class woman educated in the USA, uses her business as a vehicle to empower young girls.She gives them both professional and business training, as well as microcredit [2].Other examples include a woman whose husband is no longer supporting her, who breeds rabbits to fund the education of her adolescent daughters, who also work on the rabbit farm, and a widowed market vendor who is the sole provider for her own six children and three other dependants who are orphans of relatives [3].These cases show how entrepreneurship has enabled women to engage in feminist action beyond the subsistence of themselves and their children [4].

Women entrepreneurs in Sweden
In the case of Sweden, gendered constraints in entrepreneurship also emerged as a dominant aggregate dimension in the interviews (Table 2).Business as a tool for social change was another theme, although less dominant.In contrast to Tanzania, the feminist agency was not a dimension.The second-order themes are even more dissimilar from the case of Tanzania than are the aggregate dimensions.The representative quotes are presented in Table 2.In the text below, we have chosen one quote to illustrate each second-order theme.
4.2.1 Gendered constraints in entrepreneurship.The formal obstacles concerned the contractual conditions in the female gender-labelled health and care sector.Not only do public procurement procedures for running elderly care homes, for example, require larger organisations but the contract periods are also very short [5].
We have received 6 months at a time, and to be able to find more opportunities for this target group, who are disabled [. ..]I have had lots of ideas, but consequently not been able to pursue them.Barbara As municipalities realised that the public procurement procedures did not favour SMEs as intended, new systems such as customer-choice models were introduced.However, these models also required a providing organisation to have large resources to be able to navigate and compete strategically and successfully and place the right bids on the right units, and so on [6].
Sweden might be one of the most equal countries in the world, but there are still gendered norms constraining women as entrepreneurs, which was clear from the analysis of the interviews.Table 2 highlights some of the quotes where women talk about the attitudes that they meet from business partners and employees.While, in contrast to Tanzania, open discrimination is not an issue, nor are there widely acknowledged issues of domestic violence, there are still issues with, among other things, legitimacy.
It is still the case that if you meet an older man, he'll say 'Oh, you are the CEO -I thought you were the secretary.'Well, that's the kind of thing we hear.Maria While policy-makers have portrayed success stories of women entrepreneurs in the era of public sector downsizing in Sweden, our analysis reveals disempowerment, rather than empowerment, as a theme.In one mid-sized municipality, Siv, who owns an elderly care home in a mid-sized municipality, found it impossible to maintain the professional quality and home-like atmosphere which was her business idea.She cannot compete against the multinational corporations and has given up hope.
At first, it was good to have these small [firms] but then gradually the large ones come in and then you are pushed out.[. ..]No, I have put in my last bid.I won't bid again [. ..]I have not received anything in seven years, so there is no point.Siv In another municipality, four nurses started a care home at the beginning of the 1990s.They ran it for ten years with good results in terms of quality, satisfied users and excellent cooperation with the municipality.The nurses had offers to sell to larger companies but refused.When we met them a few years into the new millennium, they had lost the contract to one of the large companies by the lowest possible margin.Of the four nurses, one had passed away, one was on long-term sick leave due to over-work, and the remaining two were unemployed.The latter were not interested in working in elderly care or with the municipality ever again.
4.2.2Social change through entrepreneurship.On the brighter side, the analysis also revealed that some women entrepreneurs saw their business as a means to enable social change, in Sweden too.In contrast to Tanzania, our data did not include any cases where social change was related to gender equality.Rather, the women took on a high level of social and societal responsibility for their target groups, i.e. those receiving care, striving to improve the services for their benefit.This was the case in the residential home for children and young people who could not live in their home environment, as illustrated in the quote

Gendered effects
from Katarina (Table 2).The same applied to the elderly care entrepreneur (Barbara) and the yoga instructor (Lisa) respectively.The midwife interviewed (Erika, Table 2) was working both in the regional hospital and with her own business.She found initiating change in the maternity care in the regional organisation difficult and decided to initiate change via the mothers themselves, and hence started a business where she could meet them on her own terms.
[. ..] and when I found yoga, I was home.And then I wanted to pass on to others what I had been fortunate to receiveand to do that I had to start my own business".Lisa In sum, our analysis shows that women entrepreneurs in the era of public sector downsizing in Sweden have made important efforts and contributions.However, the stories do not convey any messages of an improved position for women in society or for the feminist agency.Rather, it is primarily a story of constraints in various contextual dimensions leading to the destruction of creativity.4.2.3Discussion: the irony of entrepreneurialism.During the time period for this study, i.e. 1997-2012, the temporal dimension of context was characterised by the implementation of neo-liberal agendas in both countries.However, the institutional, social and spatial contexts (Welter et al., 2014) into which the neo-liberal agenda was implemented were completely different.This meant that the resulting institutional reforms differed (Author et al., forthcoming) and that the consequences for the position of women in society and their ability to take feminist action also differed.On an overall level, the constraints experienced by the entrepreneurs mirror these differences.
Based on our interview analysis we can confirm the propositions of Author et al. (forthcoming) which suggested that in the Swedish context entrepreneurialism was not a step forward for the position of women entrepreneurs in society.Neither did entrepreneurialism nor cuts in the public sector promote women's ability to take feminist action.There were cases in our empirical material where the enterprises were used as vehicles for social change, but not patently feminist social change.Loss of income, creativity and energy were more common stories in our material, as demonstrated in previous studies (Sundin and Rapp, 2006;Sundin, 2011).Women entrepreneurs went from low-wage labour to low-income business ownership.The shift has been from a situation where the temporal, institutional, social and spatial dimensions of context were all relatively conducive for women in general, and hence also for the women who were entrepreneurs.Sweden had a welfare state, where women with a high level of trust in fair institutions and markets ventured into entrepreneurship careers that crashed because markets were not as fair and equal as they expected; rather, the privileged men and were geared towards large (maleowned) businesses.The women entrepreneurs experienced a destruction of creativity rather than creative destruction (Schumpeter, 1934(Schumpeter, /1983;;Sundin and Tillmar, 2008).From the perspective of women entrepreneurs, the Swedish case hence illustrates the dark sides of neo-liberalism and entrepreneurialism (Wright and Zahra, 2011).
In Tanzania, on the other hand, there is no trusted welfare state that redistributes resources to women, children and the elderly.As to the institutional context, the level of corruption and lack of resources was reflected in the lack of trust for the institutions.The gendered social context which also favoured men in business relations was taken for granted by the interviewees.There were no expectations whatsoever among the interviewed women that any institution, organisation or relationship would be gender-neutral.In Welter et al.'s (2014) terms, the physical space of urban Tanzania offers a much smaller place for women entrepreneurs than does urban Sweden.Welfare provision, another part of the institutional context, was seen as the responsibility of family and kin.A social context categorised by patriarchal norms also implied unpaid labour by women.Hence, the place for women's entrepreneurship and feminist agency was constrained.
Entrepreneurialism is clearly not a miracle medicine for gender equality in Tanzania.Our results are in line with previous studies from Tanzania (Langevang et al., 2018), as well as Nigeria (Sajuyigbe and Fadeyibi, 2017) and Zimbabwe (Mazonde and Carmichael, 2016) with regard to the numerous gendered socio-cultural constraints.Our analysis showed how the gendering of the institutional, social and spatial contexts was reproduced in the world of entrepreneurship, in marketplaces, in banks and in the judicial system.Another salient theme in the analysis was how men strived to conserve the gender order within the home by means of intimate partner violence against the women entrepreneurs.Within social science and medicine, this is a well-known phenomenon (Vyas et al., 2015), yet this obstacle is rarely discussed explicitly in the entrepreneurship literature.
At the same time, a strong theme was the positive effects on the living standards and opportunities for feminist agency for many women entrepreneurs.Business ownership among women has become socially accepted and it has given women opportunities to support themselves and their children independently of the patriarchal social context.Compared to the temporal context (Welter, 2011;de Bruin et al., 2014) prior to entrepreneurialism, the room for feminist agency among urban women entrepreneurs appears to have increased.We analysed cases where women used the enterprise form to improve their own position and take feminist action for other women, i.e. cases of FemInc.Ism (Ahl et al., 2016).The collective women's support groups -Chamaswere, for example, used creatively for entrepreneurship support in relation to overcoming gendered obstacles, in line with previous studies highlighting women's agency (Kinyanjui, 2012).The Tanzanian case hence illustrated acomparativelybrighter side of neo-liberalism and entrepreneurialism, at least from the perspective of urban women entrepreneurs.
To illustrate the complexity of the contextual dimensions we propose a theoretical model which has been informed by Welter et al. (2014), as well as by the current analysis (Figure 1).
Like Welter et al. (2014), we propose a model of context where gender is placed at the centre, impacting every other dimension.However, while Welter et al. illustrated the four dimensions of context on an equal basis, we argue that the temporal context should be seen Gendered effects as harbouring the whole process of social, spatial and institutional interaction.This can be interpreted to mean that the institutional, social and spatial interaction will be different at different times, but in the current analysis it is actually the other way around: the same temporal dimension (entrepreneurialism) resulted in different reforms in different, and differently gendered, institutional, social and spatial contexts.Because of the differences in these dimensions between the two countries, the effects for women entrepreneurs also differed.Figure 2 summarises the effects for women entrepreneurs in the two contexts.The double-headed arrows used in the model also illustrate, in line with the paradox of embedded agency, that women's entrepreneurship, in turn, affects the context, not the least its gender dimension.For example, women's increased independence through entrepreneurship in Tanzania may affect women's position in society.

Conclusions and contributions
In this paper, we have contributed to entrepreneurship studies in three ways.Firstly, we add to the understanding of how and why contexts matter and are differently gendered.We propose a model illustrating the dimensions of context in which we place gender at the centre.Gender influences and is influenced by the social, spatial and institutional contexts, and all these dimensions are harboured by the temporal context (Figure 1).We also develop a model illustrating how these contextual dimensions impact women entrepreneurs and their agency (Figure 2).
Secondly, by contrasting a well-functioning welfare state in the global North with a developing state in the Global Southon the immense and underexplored African continentwe broaden the empirical base of spaces studied when theorising gendering and entrepreneurship.
Thirdly, we challenge existing assumptions on the benefits of entrepreneurialism.In these neo-liberal times, it is frequently taken for granted that increased business ownership among womenand privatisation in female-dominated industriesimproves not only economic development but also gender equality.By analysing two contrasting countries affected by entrepreneurialism, we show that such a development is not necessarily the case.While we do not take a stance in the discussion of whether women entrepreneurs are "untapped resources" for economic development (Ramadani et al., 2013;Ratten and Jones, 2018), our analysis shows that entrepreneurialism far from always improves the position of women themselves.Based on our empirical analysis, we can confirm that "how neoliberalism impacts pre-conditions for women entrepreneurs is highly dependent on the local institutional framework in terms of a trustworthy women-friendly state and gender equality" (Author et al., forthcoming, p. X).While entrepreneurship can be seen as a means to give women independence and wealth, it can also be seen as a new way of exploiting women and maintaining their subordination.It depends on which alternatives are available in the context at hand.The analysis further demonstrates that reforms must be analysed through all dimensions of context, including gender, if the consequences are to be understood.
Ultimately, our analysis amounts to a criticism of neo-liberalism in general and entrepreneurialism in particular, also from an entrepreneurship perspective.Neo-liberal entrepreneurialism, stemming from the Anglo-Saxon world, was not a success in our case of a functioning state with well-aligned and trusted formal and informal institutions; at least, not for the position of women entrepreneurs and feminist agency.Only in our example of a poor developing country in the global south, struggling with institutional malfunction, could we see positive effects for the position of women entrepreneurs and feminist agency.

Policy implications
If policy measures are to support not only economic development but also gender equality, our results indicate the importance of valuing and protecting a women-friendly welfare state such as in Sweden, and of carefully analysing the value of market solutions for welfare provision.State-owned welfare services may simply be better for gender equality.The Tanzanian case, however, shows that a scenario with a women-friendly welfare state requires a different institutional and social context than what is at hand.Based on this paper, and Author et al.
(forthcoming), we argue that such a scenario is not realistic until there is an economically independent state, which is well-integrated with local norms and traditions, as well as a certain level of economic development and well-being in the community.Otherwise, the result, as in Tanzania, is likely to be corruption and subsequent distrust.Given that Tanzania does not have a women-friendly welfare state, we argue that entrepreneurship on the market may currently be more, or at least equally, conducive to feminist agency than the state.Yet, as the analysis also showed the reproduction of gendered discrimination in entrepreneurship also in Tanzania, we cannot conclude that market solutions will suffice.Entrepreneurship alone will not enable gender equality.Based on our studies, we argue that a multi-level approach to feminist change is necessary.

Further studies
As Welter et al. (2014) recognise, the dimensions of context are highly intertwined and interdependent and must be analysed holistically.In mainstream entrepreneurship studies, there is a focus on the institutional context (Bosma et al., 2008;Klofsten et al., 2016).From our analysis, it became apparent that the social and spatial context can have severe material and economic consequences for entrepreneurship.The norms and values that discriminate against women entrepreneurs in Tanzania lead to domestic violence and destroyed and cancelled business activities.
The place for women in entrepreneurship was constrained in both countries, although the gendered constraints were less materially visible in the physical space in Sweden than in Tanzania.Women entrepreneurs in Sweden met obstacles, which destroyed their creativity and motivationnot their property and bodies.Hence, a further analysis of how contexts are differently gendered could contrast also material and immaterial consequences in different spatial and temporal contexts.
As the traditional kinship structures are stronger in the rural areas in Tanzanian society, and the economy is agriculture-based, there is reason to believe that the entrepreneurialism Gendered effects promoted by the neo-liberal policies has different effects on women (and men) in rural areas than in urban (Havnevik, 2015;Otto and Ståhl, 2015).In Sweden, gender roles are often more traditional in rural areas than in urban areas, and there has been an exodus of young women from rural to urban Sweden (Westholm and Waldenström, 2008).Exploring the urban-rural dimension in the field of women's entrepreneurship would hence also be an interesting avenue for further studies.
Our critique of entrepreneurialism gives rise to a number of questions for further studies on the gendering of contexts, not least in relation to "the market".What happened in Sweden was that the market was not as free as the entrepreneurs had expected.Contrary to their expectations, the entrepreneurial opportunities were heavily skewed towards big business.In Tanzania, there was no expectation that the market would be free from the gendered institutional context, which is why the entrepreneurs did not encounter unexpected challenges.While the role of context for entrepreneurship is increasingly acknowledged, the idea that the market is always a liberating force is rarely questioned.Nonetheless, the analysis in this paper gives rise to important questions on markets in relation to gender discrimination.

Figure 1 .
Figure 1.The dimensions of contexts

Figure 2 .
Figure 2. Contextual effects for women entrepreneurs My husband died but I have struggled to put my children [my daughter and her little brother] through [private] school right up to university" (Patricia)"My own knowledge is very different now [that I am running my business alone] from when I was married.It is better to be self-dependent.You think better alone and are more clever.Before, I had to think about him, and he had to agree to truly everything" (Veronica) Schoolgirls from Form Six.They work for me and they learn, you see.I train them and eventually I end up helping them out with tuition, school fees, food or whatever" (Jasmine) "If you are not married, they are not afraid of you at all, they can do anything" (Barika) Lack of respect for single womenHusband left to live with second wife due to the business (Fatima) "Women are being beaten. . ..because not all men want a woman to have access to finance in society.Of course, women are trying hard, but those are the issues that we face and as coaches and as mentors we sometimes have to carry the burden of those women because we receive their calls for [0..] . . .you know . . .giving them courage" (Karen) "I was in the batik business.He started to scream and fight.So I stopped that business and started with vegetables.The business grew a bit, then he started big fights and destroyed everything in the field and destroyed the vegetables" (Mary) "We can help each other.To give each other information, for example, if there are exhibitions somewhere everyone can get the information.The 'chama' is very lively" .(Barika)"I have started a group with people from here.We give each other advise and we have started to lend each other small sums of money.We meet every Saturday from 9 to 10" (Mary) "All women have problems, if you tell others you will see that your own [worries] decrease" (Mary) "I think it will be very difficult for small businesses . . .The competition is not fair.There are no small businesses that can place zero-sum bids, to take a chance and see how it goes.You have to know what you've got" (Ingegerd) "It is still the case that if you meet an older man, he'll say 'Oh, you are the CEO -I thought you were the secretary.'Well, that¨s the kind of thing we hear" (Maria) "It's like it is in the air, and then if you need to be someone's boss as well, then . . .well, they do not accept that you are their superior" (Maja) "Finally, you get so tired you say 'I don't give a shit'.Perhaps we should have an old man making the decisions around here . . .then everything would be fine . . .That is exactly where I didn't want to end up, what I have been fighting for since I was a teenager, that I would be in the situation where I have been through it all and am back where I started" (Maja)."I have felt [the attitude] that kind of 'can she, that girl, really do this?'" (Barbara) "It doesn't help that I have lots of energy and willingness and that I bring up clear ideas . . .We are not getting anywhere.So now, I say that I have used up my stock of willingness" (Barbara) "At first, it was good to have these small [firms] but then gradually the large ones come in and then you are pushed out. . . .No, I have put in my last bid.I won't bid again . . .I have not received anything in seven years, so there is no point" (Siv) "We have exhausted it, if you like, we haven't the energy to do it all over again.I will never work for the municipality ever again" (Margret) sectors "We have a pedagogical approach which is very unusual . . .we do not regard these kids as sick . . .we have a burning interest in these kids and what we can do to help them" (Katarina) "The most important thing is that the target group has some security . ... I do not want to leave when there is a crisis.I want to see it carried on.It is that sense of responsibility one has" (Barbara)